| Popis: |
Bone mineral density (BMD) is a highly heritable trait, indicating that genetic elements are partly responsible for variation in osteoporosis risk. To further understand the genetic variation underlying osteoporosis, I performed genome-wide association (GWA) studies using designs that have largely not been performed in osteoporosis literature to date. Three strategies were used: i) a selected sample of postmenopausal women with high (z ~ 1.5, n = 1,055) or low (z≥1.5, n = 900) hip BMD [as measured by Dual-energy X-ray absorptiometry (DXA)] were used for GWA, followed by replication in an unselected sample of 20,898 adults, ii) a GW A meta-analysis on unselected children from the Avon Longitudinal Study of Parents and Children [ALSPAC (n = 5,330)] and the Generation R study [GEN-R (n = 4,098)], using DXA derived total-body less head BMD (TBLH-BMD) measures, iii) refining total-body BMD measures in children by subregional analysis: i.e. quantifying the genetic and environmental correlation between paediatric BMD measures [ALSPAC (n≥5,299)] of the skull (S-BMD), lower limb (LL-BMD) and upper limb (UL-BMD) using genome-wide complex trait analysis (GCT A) and thereafter performing a GWA meta-analysis on each site using subjects from ALSPAC and GEN-R (n ~ 9,300). The role of bone resorption in bone growth and accrual was investigated via a cross-sectional analysis of 1,130 adolescents from ALSPAC using serum measures of, β-C-telopeptides of type I collagen (CTX) and quantitative computed tomography (pQCT) measures of the mid-tibia. Two novel BMD associated loci were identified using the selective genotyping strategy: GALNTJ (rs6710518, P = 1.4x lO· 10) and RSP03 (rsI3204965, P = 3.0x lO· 10). Association studies of paediatric TBLH-BMD identified a novel variant in RlN3 (rs754388, P = 3.0x 10.9) and replicated 31 adult BMD associated loci, with six reaching the GWA threshold of association (P < 5.0x 10·R). Sub-regional GCT A analysis indicated that appendicular sites shared a greater proportion of genetic architecture (LL-/UL-BMD rg=0.78, P = 1 x 10.7) when compared to the skull [(UL-/SBMD rg = 0.58, P = 9x l0·7) and (LL-/S-BMD rg = 0.43, P = l x lO'~)]. GWA meta-analysis echoed these findings by identifying twelve known BMD-associated variants that differed in the strength of their association and magnitude of effect with each sub-region. In particular, variants at the WNTl6 and RSP03 showed considerable site-specificity as indicated by strong association with S-BMD and/or UL-BMD, but not with LL-BMD. An investigation into the role of bone resorption in adolescent bone suggested that CTX was positively related to periosteal circumference (PC) [,8 = 0.19 (0 .13, 0.24)] (coefficient = SD change per SD increase in CTX, 95% Cl)], but inversely associated with cortical BMD [,8 = -0.46 (-0.52, -0.40)] and positively related to bone strength as reflected by the strength-strain index (SSI) [,8 = 0.09 (0 .03, 0.14)]. These relationships were replicated using genetic proxies for bone resorption. . These results suggest that the selective sampling GWA strategy represents an efficient alternative to conventional random sampling designs. However the real world feasibility of selective sampling is questionable, as it requires extensive phenotyping in order to ensure adequate sample size and study power is obtained. BMD measures of children are well suited for GW A, however the replication of adult BMD associated SNPs implies that many of the BMD associated loci identified operate throughout the life course. Whether this strategy enriches for genetic factors involved in bone modelling remains to be seen. BMD at different skeletal sites appears to be influenced by distinct genetic and environmental influences, suggesting that phenotypic refinement of BMD may represent a superior GW A strategy, when compared to using heterogeneous BMO measures (i .e TBLH-BMO). Finally, bone resorption might play an important role in paediatric bone growth, accrual and strength. |
